X-Git-Url: https://code.delx.au/gnu-emacs/blobdiff_plain/d02fe47dd3be7310d1bfd6e802d1fac2ea5f5e9d..95b1abcfafe8a366a75635f5fa4b4fa1e79f2964:/src/search.c
diff --git a/src/search.c b/src/search.c
index 96daecb728..2269afc6d8 100644
--- a/src/search.c
+++ b/src/search.c
@@ -1,14 +1,14 @@
/* String search routines for GNU Emacs.
Copyright (C) 1985, 1986, 1987, 1993, 1994, 1997, 1998, 1999, 2001, 2002,
- 2003, 2004, 2005, 2006, 2007, 2008
+ 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010
Free Software Foundation, Inc.
This file is part of GNU Emacs.
-GNU Emacs is free software; you can redistribute it and/or modify
+GNU Emacs is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
-the Free Software Foundation; either version 3, or (at your option)
-any later version.
+the Free Software Foundation, either version 3 of the License, or
+(at your option) any later version.
GNU Emacs is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
@@ -16,12 +16,11 @@ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
-along with GNU Emacs; see the file COPYING. If not, write to
-the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
-Boston, MA 02110-1301, USA. */
+along with GNU Emacs. If not, see . */
#include
+#include
#include "lisp.h"
#include "syntax.h"
#include "category.h"
@@ -99,11 +98,18 @@ Lisp_Object Vsearch_spaces_regexp;
only. */
Lisp_Object Vinhibit_changing_match_data;
-static void set_search_regs ();
-static void save_search_regs ();
-static int simple_search ();
-static int boyer_moore ();
-static int search_buffer ();
+static void set_search_regs P_ ((EMACS_INT, EMACS_INT));
+static void save_search_regs P_ ((void));
+static EMACS_INT simple_search P_ ((int, unsigned char *, int, int,
+ Lisp_Object, EMACS_INT, EMACS_INT,
+ EMACS_INT, EMACS_INT));
+static EMACS_INT boyer_moore P_ ((int, unsigned char *, int, int,
+ Lisp_Object, Lisp_Object,
+ EMACS_INT, EMACS_INT,
+ EMACS_INT, EMACS_INT, int));
+static EMACS_INT search_buffer P_ ((Lisp_Object, EMACS_INT, EMACS_INT,
+ EMACS_INT, EMACS_INT, int, int,
+ Lisp_Object, Lisp_Object, int));
static void matcher_overflow () NO_RETURN;
static void
@@ -204,8 +210,8 @@ clear_regexp_cache ()
int i;
for (i = 0; i < REGEXP_CACHE_SIZE; ++i)
- /* It's tempting to compare with the syntax-table we've actually changd,
- but it's not sufficient because char-table inheritance mewans that
+ /* It's tempting to compare with the syntax-table we've actually changed,
+ but it's not sufficient because char-table inheritance means that
modifying one syntax-table can change others at the same time. */
if (!EQ (searchbufs[i].syntax_table, Qt))
searchbufs[i].regexp = Qnil;
@@ -290,7 +296,7 @@ looking_at_1 (string, posix)
{
Lisp_Object val;
unsigned char *p1, *p2;
- int s1, s2;
+ EMACS_INT s1, s2;
register int i;
struct re_pattern_buffer *bufp;
@@ -392,7 +398,7 @@ string_match_1 (regexp, string, start, posix)
{
int val;
struct re_pattern_buffer *bufp;
- int pos, pos_byte;
+ EMACS_INT pos, pos_byte;
int i;
if (running_asynch_code)
@@ -558,6 +564,74 @@ fast_string_match_ignore_case (regexp, string)
immediate_quit = 0;
return val;
}
+
+/* Match REGEXP against the characters after POS to LIMIT, and return
+ the number of matched characters. If STRING is non-nil, match
+ against the characters in it. In that case, POS and LIMIT are
+ indices into the string. This function doesn't modify the match
+ data. */
+
+EMACS_INT
+fast_looking_at (regexp, pos, pos_byte, limit, limit_byte, string)
+ Lisp_Object regexp;
+ EMACS_INT pos, pos_byte, limit, limit_byte;
+ Lisp_Object string;
+{
+ int multibyte;
+ struct re_pattern_buffer *buf;
+ unsigned char *p1, *p2;
+ EMACS_INT s1, s2;
+ EMACS_INT len;
+
+ if (STRINGP (string))
+ {
+ if (pos_byte < 0)
+ pos_byte = string_char_to_byte (string, pos);
+ if (limit_byte < 0)
+ limit_byte = string_char_to_byte (string, limit);
+ p1 = NULL;
+ s1 = 0;
+ p2 = SDATA (string);
+ s2 = SBYTES (string);
+ re_match_object = string;
+ multibyte = STRING_MULTIBYTE (string);
+ }
+ else
+ {
+ if (pos_byte < 0)
+ pos_byte = CHAR_TO_BYTE (pos);
+ if (limit_byte < 0)
+ limit_byte = CHAR_TO_BYTE (limit);
+ pos_byte -= BEGV_BYTE;
+ limit_byte -= BEGV_BYTE;
+ p1 = BEGV_ADDR;
+ s1 = GPT_BYTE - BEGV_BYTE;
+ p2 = GAP_END_ADDR;
+ s2 = ZV_BYTE - GPT_BYTE;
+ if (s1 < 0)
+ {
+ p2 = p1;
+ s2 = ZV_BYTE - BEGV_BYTE;
+ s1 = 0;
+ }
+ if (s2 < 0)
+ {
+ s1 = ZV_BYTE - BEGV_BYTE;
+ s2 = 0;
+ }
+ re_match_object = Qnil;
+ multibyte = ! NILP (current_buffer->enable_multibyte_characters);
+ }
+
+ buf = compile_pattern (regexp, 0, Qnil, 0, multibyte);
+ immediate_quit = 1;
+ len = re_match_2 (buf, (char *) p1, s1, (char *) p2, s2,
+ pos_byte, NULL, limit_byte);
+ immediate_quit = 0;
+
+ return len;
+}
+
/* The newline cache: remembering which sections of text have no newlines. */
@@ -610,7 +684,7 @@ newline_cache_on_off (buf)
int
scan_buffer (target, start, end, count, shortage, allow_quit)
register int target;
- int start, end;
+ EMACS_INT start, end;
int count;
int *shortage;
int allow_quit;
@@ -645,9 +719,9 @@ scan_buffer (target, start, end, count, shortage, allow_quit)
the position of the last character before the next such
obstacle --- the last character the dumb search loop should
examine. */
- int ceiling_byte = CHAR_TO_BYTE (end) - 1;
- int start_byte = CHAR_TO_BYTE (start);
- int tem;
+ EMACS_INT ceiling_byte = CHAR_TO_BYTE (end) - 1;
+ EMACS_INT start_byte = CHAR_TO_BYTE (start);
+ EMACS_INT tem;
/* If we're looking for a newline, consult the newline cache
to see where we can avoid some scanning. */
@@ -718,9 +792,9 @@ scan_buffer (target, start, end, count, shortage, allow_quit)
while (start > end)
{
/* The last character to check before the next obstacle. */
- int ceiling_byte = CHAR_TO_BYTE (end);
- int start_byte = CHAR_TO_BYTE (start);
- int tem;
+ EMACS_INT ceiling_byte = CHAR_TO_BYTE (end);
+ EMACS_INT start_byte = CHAR_TO_BYTE (start);
+ EMACS_INT tem;
/* Consult the newline cache, if appropriate. */
if (target == '\n' && newline_cache)
@@ -806,8 +880,8 @@ scan_buffer (target, start, end, count, shortage, allow_quit)
int
scan_newline (start, start_byte, limit, limit_byte, count, allow_quit)
- int start, start_byte;
- int limit, limit_byte;
+ EMACS_INT start, start_byte;
+ EMACS_INT limit, limit_byte;
register int count;
int allow_quit;
{
@@ -816,7 +890,7 @@ scan_newline (start, start_byte, limit, limit_byte, count, allow_quit)
register unsigned char *cursor;
unsigned char *base;
- register int ceiling;
+ EMACS_INT ceiling;
register unsigned char *ceiling_addr;
int old_immediate_quit = immediate_quit;
@@ -904,7 +978,8 @@ scan_newline (start, start_byte, limit, limit_byte, count, allow_quit)
int
find_next_newline_no_quit (from, cnt)
- register int from, cnt;
+ EMACS_INT from;
+ int cnt;
{
return scan_buffer ('\n', from, 0, cnt, (int *) 0, 0);
}
@@ -915,7 +990,8 @@ find_next_newline_no_quit (from, cnt)
int
find_before_next_newline (from, to, cnt)
- int from, to, cnt;
+ EMACS_INT from, to;
+ int cnt;
{
int shortage;
int pos = scan_buffer ('\n', from, to, cnt, &shortage, 1);
@@ -1076,14 +1152,14 @@ while (0)
(i.e. Vinhibit_changing_match_data is non-nil). */
static struct re_registers search_regs_1;
-static int
+static EMACS_INT
search_buffer (string, pos, pos_byte, lim, lim_byte, n,
RE, trt, inverse_trt, posix)
Lisp_Object string;
- int pos;
- int pos_byte;
- int lim;
- int lim_byte;
+ EMACS_INT pos;
+ EMACS_INT pos_byte;
+ EMACS_INT lim;
+ EMACS_INT lim_byte;
int n;
int RE;
Lisp_Object trt;
@@ -1311,7 +1387,7 @@ search_buffer (string, pos, pos_byte, lim, lim_byte, n,
base_pat++;
}
- c = STRING_CHAR_AND_LENGTH (base_pat, len_byte, in_charlen);
+ c = STRING_CHAR_AND_LENGTH (base_pat, in_charlen);
if (NILP (trt))
{
@@ -1342,11 +1418,7 @@ search_buffer (string, pos, pos_byte, lim, lim_byte, n,
if (this_char_base > 0)
boyer_moore_ok = 0;
else
- {
- this_char_base = 0;
- if (char_base < 0)
- char_base = this_char_base;
- }
+ this_char_base = 0;
}
else if (CHAR_BYTE8_P (inverse))
/* Boyer-moore search can't handle a
@@ -1358,8 +1430,7 @@ search_buffer (string, pos, pos_byte, lim, lim_byte, n,
this_char_base = inverse & ~0x3F;
if (char_base < 0)
char_base = this_char_base;
- else if (char_base > 0
- && this_char_base != char_base)
+ else if (this_char_base != char_base)
boyer_moore_ok = 0;
}
else if ((inverse & ~0x3F) != this_char_base)
@@ -1370,8 +1441,6 @@ search_buffer (string, pos, pos_byte, lim, lim_byte, n,
}
}
}
- if (char_base < 0)
- char_base = 0;
/* Store this character into the translated pattern. */
bcopy (str, pat, charlen);
@@ -1379,6 +1448,11 @@ search_buffer (string, pos, pos_byte, lim, lim_byte, n,
base_pat += in_charlen;
len_byte -= in_charlen;
}
+
+ /* If char_base is still negative we didn't find any translated
+ non-ASCII characters. */
+ if (char_base < 0)
+ char_base = 0;
}
else
{
@@ -1429,14 +1503,14 @@ search_buffer (string, pos, pos_byte, lim, lim_byte, n,
regardless of what is in TRT. It is used in cases where
boyer_moore cannot work. */
-static int
+static EMACS_INT
simple_search (n, pat, len, len_byte, trt, pos, pos_byte, lim, lim_byte)
int n;
unsigned char *pat;
int len, len_byte;
Lisp_Object trt;
- int pos, pos_byte;
- int lim, lim_byte;
+ EMACS_INT pos, pos_byte;
+ EMACS_INT lim, lim_byte;
{
int multibyte = ! NILP (current_buffer->enable_multibyte_characters);
int forward = n > 0;
@@ -1450,10 +1524,9 @@ simple_search (n, pat, len, len_byte, trt, pos, pos_byte, lim, lim_byte)
while (1)
{
/* Try matching at position POS. */
- int this_pos = pos;
- int this_pos_byte = pos_byte;
+ EMACS_INT this_pos = pos;
+ EMACS_INT this_pos_byte = pos_byte;
int this_len = len;
- int this_len_byte = len_byte;
unsigned char *p = pat;
if (pos + len > lim || pos_byte + len_byte > lim_byte)
goto stop;
@@ -1463,16 +1536,14 @@ simple_search (n, pat, len, len_byte, trt, pos, pos_byte, lim, lim_byte)
int charlen, buf_charlen;
int pat_ch, buf_ch;
- pat_ch = STRING_CHAR_AND_LENGTH (p, this_len_byte, charlen);
+ pat_ch = STRING_CHAR_AND_LENGTH (p, charlen);
buf_ch = STRING_CHAR_AND_LENGTH (BYTE_POS_ADDR (this_pos_byte),
- ZV_BYTE - this_pos_byte,
buf_charlen);
TRANSLATE (buf_ch, trt, buf_ch);
if (buf_ch != pat_ch)
break;
- this_len_byte -= charlen;
this_len--;
p += charlen;
@@ -1499,7 +1570,7 @@ simple_search (n, pat, len, len_byte, trt, pos, pos_byte, lim, lim_byte)
while (1)
{
/* Try matching at position POS. */
- int this_pos = pos;
+ EMACS_INT this_pos = pos;
int this_len = len;
unsigned char *p = pat;
@@ -1538,42 +1609,36 @@ simple_search (n, pat, len, len_byte, trt, pos, pos_byte, lim, lim_byte)
while (1)
{
/* Try matching at position POS. */
- int this_pos = pos - len;
- int this_pos_byte;
+ EMACS_INT this_pos = pos;
+ EMACS_INT this_pos_byte = pos_byte;
int this_len = len;
- int this_len_byte = len_byte;
- unsigned char *p = pat;
+ const unsigned char *p = pat + len_byte;
- if (this_pos < lim || (pos_byte - len_byte) < lim_byte)
+ if (this_pos - len < lim || (pos_byte - len_byte) < lim_byte)
goto stop;
- this_pos_byte = CHAR_TO_BYTE (this_pos);
- match_byte = pos_byte - this_pos_byte;
while (this_len > 0)
{
- int charlen, buf_charlen;
+ int charlen;
int pat_ch, buf_ch;
- pat_ch = STRING_CHAR_AND_LENGTH (p, this_len_byte, charlen);
- buf_ch = STRING_CHAR_AND_LENGTH (BYTE_POS_ADDR (this_pos_byte),
- ZV_BYTE - this_pos_byte,
- buf_charlen);
+ DEC_BOTH (this_pos, this_pos_byte);
+ PREV_CHAR_BOUNDARY (p, pat);
+ pat_ch = STRING_CHAR (p);
+ buf_ch = STRING_CHAR (BYTE_POS_ADDR (this_pos_byte));
TRANSLATE (buf_ch, trt, buf_ch);
if (buf_ch != pat_ch)
break;
- this_len_byte -= charlen;
this_len--;
- p += charlen;
- this_pos_byte += buf_charlen;
- this_pos++;
}
if (this_len == 0)
{
- pos -= len;
- pos_byte -= match_byte;
+ match_byte = pos_byte - this_pos_byte;
+ pos = this_pos;
+ pos_byte = this_pos_byte;
break;
}
@@ -1588,7 +1653,7 @@ simple_search (n, pat, len, len_byte, trt, pos, pos_byte, lim, lim_byte)
while (1)
{
/* Try matching at position POS. */
- int this_pos = pos - len;
+ EMACS_INT this_pos = pos - len;
int this_len = len;
unsigned char *p = pat;
@@ -1651,7 +1716,7 @@ simple_search (n, pat, len, len_byte, trt, pos, pos_byte, lim, lim_byte)
If that criterion is not satisfied, do not call this function. */
-static int
+static EMACS_INT
boyer_moore (n, base_pat, len, len_byte, trt, inverse_trt,
pos, pos_byte, lim, lim_byte, char_base)
int n;
@@ -1659,15 +1724,15 @@ boyer_moore (n, base_pat, len, len_byte, trt, inverse_trt,
int len, len_byte;
Lisp_Object trt;
Lisp_Object inverse_trt;
- int pos, pos_byte;
- int lim, lim_byte;
+ EMACS_INT pos, pos_byte;
+ EMACS_INT lim, lim_byte;
int char_base;
{
int direction = ((n > 0) ? 1 : -1);
register int dirlen;
- int infinity, limit, stride_for_teases = 0;
- register int *BM_tab;
- int *BM_tab_base;
+ EMACS_INT limit;
+ int stride_for_teases = 0;
+ int BM_tab[0400];
register unsigned char *cursor, *p_limit;
register int i, j;
unsigned char *pat, *pat_end;
@@ -1683,37 +1748,28 @@ boyer_moore (n, base_pat, len, len_byte, trt, inverse_trt,
int translate_prev_byte3 = 0;
int translate_prev_byte4 = 0;
- BM_tab = (int *) alloca (0400 * sizeof (int));
-
- /* The general approach is that we are going to maintain that we know */
- /* the first (closest to the present position, in whatever direction */
- /* we're searching) character that could possibly be the last */
- /* (furthest from present position) character of a valid match. We */
- /* advance the state of our knowledge by looking at that character */
- /* and seeing whether it indeed matches the last character of the */
- /* pattern. If it does, we take a closer look. If it does not, we */
- /* move our pointer (to putative last characters) as far as is */
- /* logically possible. This amount of movement, which I call a */
- /* stride, will be the length of the pattern if the actual character */
- /* appears nowhere in the pattern, otherwise it will be the distance */
- /* from the last occurrence of that character to the end of the */
- /* pattern. */
- /* As a coding trick, an enormous stride is coded into the table for */
- /* characters that match the last character. This allows use of only */
- /* a single test, a test for having gone past the end of the */
- /* permissible match region, to test for both possible matches (when */
- /* the stride goes past the end immediately) and failure to */
- /* match (where you get nudged past the end one stride at a time). */
-
- /* Here we make a "mickey mouse" BM table. The stride of the search */
- /* is determined only by the last character of the putative match. */
- /* If that character does not match, we will stride the proper */
- /* distance to propose a match that superimposes it on the last */
- /* instance of a character that matches it (per trt), or misses */
- /* it entirely if there is none. */
+ /* The general approach is that we are going to maintain that we know
+ the first (closest to the present position, in whatever direction
+ we're searching) character that could possibly be the last
+ (furthest from present position) character of a valid match. We
+ advance the state of our knowledge by looking at that character
+ and seeing whether it indeed matches the last character of the
+ pattern. If it does, we take a closer look. If it does not, we
+ move our pointer (to putative last characters) as far as is
+ logically possible. This amount of movement, which I call a
+ stride, will be the length of the pattern if the actual character
+ appears nowhere in the pattern, otherwise it will be the distance
+ from the last occurrence of that character to the end of the
+ pattern. If the amount is zero we have a possible match. */
+
+ /* Here we make a "mickey mouse" BM table. The stride of the search
+ is determined only by the last character of the putative match.
+ If that character does not match, we will stride the proper
+ distance to propose a match that superimposes it on the last
+ instance of a character that matches it (per trt), or misses
+ it entirely if there is none. */
dirlen = len_byte * direction;
- infinity = dirlen - (lim_byte + pos_byte + len_byte + len_byte) * direction;
/* Record position after the end of the pattern. */
pat_end = base_pat + len_byte;
@@ -1723,23 +1779,14 @@ boyer_moore (n, base_pat, len, len_byte, trt, inverse_trt,
if (direction < 0)
base_pat = pat_end - 1;
- BM_tab_base = BM_tab;
- BM_tab += 0400;
- j = dirlen; /* to get it in a register */
- /* A character that does not appear in the pattern induces a */
- /* stride equal to the pattern length. */
- while (BM_tab_base != BM_tab)
- {
- *--BM_tab = j;
- *--BM_tab = j;
- *--BM_tab = j;
- *--BM_tab = j;
- }
+ /* A character that does not appear in the pattern induces a
+ stride equal to the pattern length. */
+ for (i = 0; i < 0400; i++)
+ BM_tab[i] = dirlen;
/* We use this for translation, instead of TRT itself.
We fill this in to handle the characters that actually
occur in the pattern. Others don't matter anyway! */
- bzero (simple_translate, sizeof simple_translate);
for (i = 0; i < 0400; i++)
simple_translate[i] = i;
@@ -1764,12 +1811,10 @@ boyer_moore (n, base_pat, len, len_byte, trt, inverse_trt,
}
i = 0;
- while (i != infinity)
+ while (i != dirlen)
{
unsigned char *ptr = base_pat + i;
i += direction;
- if (i == dirlen)
- i = infinity;
if (! NILP (trt))
{
/* If the byte currently looking at is the last of a
@@ -1787,7 +1832,7 @@ boyer_moore (n, base_pat, len, len_byte, trt, inverse_trt,
while (! (CHAR_HEAD_P (*charstart)))
charstart--;
- ch = STRING_CHAR (charstart, ptr - charstart + 1);
+ ch = STRING_CHAR (charstart);
if (char_base != (ch & ~0x3F))
ch = -1;
}
@@ -1797,7 +1842,7 @@ boyer_moore (n, base_pat, len, len_byte, trt, inverse_trt,
else
j = *ptr;
- if (i == infinity)
+ if (i == dirlen)
stride_for_teases = BM_tab[j];
BM_tab[j] = dirlen - i;
@@ -1830,23 +1875,22 @@ boyer_moore (n, base_pat, len, len_byte, trt, inverse_trt,
{
j = *ptr;
- if (i == infinity)
+ if (i == dirlen)
stride_for_teases = BM_tab[j];
BM_tab[j] = dirlen - i;
}
- /* stride_for_teases tells how much to stride if we get a */
- /* match on the far character but are subsequently */
- /* disappointed, by recording what the stride would have been */
- /* for that character if the last character had been */
- /* different. */
+ /* stride_for_teases tells how much to stride if we get a
+ match on the far character but are subsequently
+ disappointed, by recording what the stride would have been
+ for that character if the last character had been
+ different. */
}
- infinity = dirlen - infinity;
pos_byte += dirlen - ((direction > 0) ? direction : 0);
/* loop invariant - POS_BYTE points at where last char (first
char if reverse) of pattern would align in a possible match. */
while (n != 0)
{
- int tail_end;
+ EMACS_INT tail_end;
unsigned char *tail_end_ptr;
/* It's been reported that some (broken) compiler thinks that
@@ -1884,43 +1928,34 @@ boyer_moore (n, base_pat, len, len_byte, trt, inverse_trt,
p_limit = BYTE_POS_ADDR (limit);
p2 = (cursor = BYTE_POS_ADDR (pos_byte));
- /* In this loop, pos + cursor - p2 is the surrogate for pos */
+ /* In this loop, pos + cursor - p2 is the surrogate for pos. */
while (1) /* use one cursor setting as long as i can */
{
if (direction > 0) /* worth duplicating */
{
- /* Use signed comparison if appropriate
- to make cursor+infinity sure to be > p_limit.
- Assuming that the buffer lies in a range of addresses
- that are all "positive" (as ints) or all "negative",
- either kind of comparison will work as long
- as we don't step by infinity. So pick the kind
- that works when we do step by infinity. */
- if ((EMACS_INT) (p_limit + infinity) > (EMACS_INT) p_limit)
- while ((EMACS_INT) cursor <= (EMACS_INT) p_limit)
- cursor += BM_tab[*cursor];
- else
- while ((EMACS_UINT) cursor <= (EMACS_UINT) p_limit)
+ while (cursor <= p_limit)
+ {
+ if (BM_tab[*cursor] == 0)
+ goto hit;
cursor += BM_tab[*cursor];
+ }
}
else
{
- if ((EMACS_INT) (p_limit + infinity) < (EMACS_INT) p_limit)
- while ((EMACS_INT) cursor >= (EMACS_INT) p_limit)
- cursor += BM_tab[*cursor];
- else
- while ((EMACS_UINT) cursor >= (EMACS_UINT) p_limit)
+ while (cursor >= p_limit)
+ {
+ if (BM_tab[*cursor] == 0)
+ goto hit;
cursor += BM_tab[*cursor];
+ }
}
-/* If you are here, cursor is beyond the end of the searched region. */
-/* This can happen if you match on the far character of the pattern, */
-/* because the "stride" of that character is infinity, a number able */
-/* to throw you well beyond the end of the search. It can also */
-/* happen if you fail to match within the permitted region and would */
-/* otherwise try a character beyond that region */
- if ((cursor - p_limit) * direction <= len_byte)
- break; /* a small overrun is genuine */
- cursor -= infinity; /* large overrun = hit */
+ /* If you are here, cursor is beyond the end of the
+ searched region. You fail to match within the
+ permitted region and would otherwise try a character
+ beyond that region. */
+ break;
+
+ hit:
i = dirlen - direction;
if (! NILP (trt))
{
@@ -1959,7 +1994,7 @@ boyer_moore (n, base_pat, len, len_byte, trt, inverse_trt,
cursor += dirlen - i - direction; /* fix cursor */
if (i + direction == 0)
{
- int position, start, end;
+ EMACS_INT position, start, end;
cursor -= direction;
@@ -1992,8 +2027,8 @@ boyer_moore (n, base_pat, len, len_byte, trt, inverse_trt,
pos_byte += cursor - p2;
}
else
- /* Now we'll pick up a clump that has to be done the hard */
- /* way because it covers a discontinuity */
+ /* Now we'll pick up a clump that has to be done the hard
+ way because it covers a discontinuity. */
{
limit = ((direction > 0)
? BUFFER_CEILING_OF (pos_byte - dirlen + 1)
@@ -2005,19 +2040,21 @@ boyer_moore (n, base_pat, len, len_byte, trt, inverse_trt,
and still be valid for a possible match. */
while (1)
{
- /* This loop can be coded for space rather than */
- /* speed because it will usually run only once. */
- /* (the reach is at most len + 21, and typically */
- /* does not exceed len) */
+ /* This loop can be coded for space rather than
+ speed because it will usually run only once.
+ (the reach is at most len + 21, and typically
+ does not exceed len). */
while ((limit - pos_byte) * direction >= 0)
- pos_byte += BM_tab[FETCH_BYTE (pos_byte)];
- /* now run the same tests to distinguish going off the */
- /* end, a match or a phony match. */
- if ((pos_byte - limit) * direction <= len_byte)
- break; /* ran off the end */
- /* Found what might be a match.
- Set POS_BYTE back to last (first if reverse) pos. */
- pos_byte -= infinity;
+ {
+ int ch = FETCH_BYTE (pos_byte);
+ if (BM_tab[ch] == 0)
+ goto hit2;
+ pos_byte += BM_tab[ch];
+ }
+ break; /* ran off the end */
+
+ hit2:
+ /* Found what might be a match. */
i = dirlen - direction;
while ((i -= direction) + direction != 0)
{
@@ -2046,10 +2083,10 @@ boyer_moore (n, base_pat, len, len_byte, trt, inverse_trt,
/* Above loop has moved POS_BYTE part or all the way
back to the first pos (last pos if reverse).
Set it once again at the last (first if reverse) char. */
- pos_byte += dirlen - i- direction;
+ pos_byte += dirlen - i - direction;
if (i + direction == 0)
{
- int position, start, end;
+ EMACS_INT position, start, end;
pos_byte -= direction;
position = pos_byte + ((direction > 0) ? 1 - len_byte : 0);
@@ -2091,7 +2128,7 @@ boyer_moore (n, base_pat, len, len_byte, trt, inverse_trt,
static void
set_search_regs (beg_byte, nbytes)
- int beg_byte, nbytes;
+ EMACS_INT beg_byte, nbytes;
{
int i;
@@ -2119,19 +2156,21 @@ set_search_regs (beg_byte, nbytes)
XSETBUFFER (last_thing_searched, current_buffer);
}
-/* Given a string of words separated by word delimiters,
- compute a regexp that matches those exact words
- separated by arbitrary punctuation. */
+/* Given STRING, a string of words separated by word delimiters,
+ compute a regexp that matches those exact words separated by
+ arbitrary punctuation. If LAX is nonzero, the end of the string
+ need not match a word boundary unless it ends in whitespace. */
static Lisp_Object
-wordify (string)
+wordify (string, lax)
Lisp_Object string;
+ int lax;
{
register unsigned char *p, *o;
register int i, i_byte, len, punct_count = 0, word_count = 0;
Lisp_Object val;
int prev_c = 0;
- int adjust;
+ int adjust, whitespace_at_end;
CHECK_STRING (string);
p = SDATA (string);
@@ -2154,11 +2193,18 @@ wordify (string)
}
if (SYNTAX (prev_c) == Sword)
- word_count++;
+ {
+ word_count++;
+ whitespace_at_end = 0;
+ }
+ else
+ whitespace_at_end = 1;
+
if (!word_count)
return empty_unibyte_string;
- adjust = - punct_count + 5 * (word_count - 1) + 4;
+ adjust = - punct_count + 5 * (word_count - 1)
+ + ((lax && !whitespace_at_end) ? 2 : 4);
if (STRING_MULTIBYTE (string))
val = make_uninit_multibyte_string (len + adjust,
SBYTES (string)
@@ -2196,8 +2242,11 @@ wordify (string)
prev_c = c;
}
- *o++ = '\\';
- *o++ = 'b';
+ if (!lax || whitespace_at_end)
+ {
+ *o++ = '\\';
+ *o++ = 'b';
+ }
return val;
}
@@ -2254,7 +2303,7 @@ Optional fourth argument is repeat count--search for successive occurrences. */
(string, bound, noerror, count)
Lisp_Object string, bound, noerror, count;
{
- return search_command (wordify (string), bound, noerror, count, -1, 1, 0);
+ return search_command (wordify (string, 0), bound, noerror, count, -1, 1, 0);
}
DEFUN ("word-search-forward", Fword_search_forward, Sword_search_forward, 1, 4,
@@ -2269,7 +2318,45 @@ Optional fourth argument is repeat count--search for successive occurrences. */
(string, bound, noerror, count)
Lisp_Object string, bound, noerror, count;
{
- return search_command (wordify (string), bound, noerror, count, 1, 1, 0);
+ return search_command (wordify (string, 0), bound, noerror, count, 1, 1, 0);
+}
+
+DEFUN ("word-search-backward-lax", Fword_search_backward_lax, Sword_search_backward_lax, 1, 4,
+ "sWord search backward: ",
+ doc: /* Search backward from point for STRING, ignoring differences in punctuation.
+Set point to the beginning of the occurrence found, and return point.
+
+Unlike `word-search-backward', the end of STRING need not match a word
+boundary unless it ends in whitespace.
+
+An optional second argument bounds the search; it is a buffer position.
+The match found must not extend before that position.
+Optional third argument, if t, means if fail just return nil (no error).
+ If not nil and not t, move to limit of search and return nil.
+Optional fourth argument is repeat count--search for successive occurrences. */)
+ (string, bound, noerror, count)
+ Lisp_Object string, bound, noerror, count;
+{
+ return search_command (wordify (string, 1), bound, noerror, count, -1, 1, 0);
+}
+
+DEFUN ("word-search-forward-lax", Fword_search_forward_lax, Sword_search_forward_lax, 1, 4,
+ "sWord search: ",
+ doc: /* Search forward from point for STRING, ignoring differences in punctuation.
+Set point to the end of the occurrence found, and return point.
+
+Unlike `word-search-forward', the end of STRING need not match a word
+boundary unless it ends in whitespace.
+
+An optional second argument bounds the search; it is a buffer position.
+The match found must not extend after that position.
+Optional third argument, if t, means if fail just return nil (no error).
+ If not nil and not t, move to limit of search and return nil.
+Optional fourth argument is repeat count--search for successive occurrences. */)
+ (string, bound, noerror, count)
+ Lisp_Object string, bound, noerror, count;
+{
+ return search_command (wordify (string, 1), bound, noerror, count, 1, 1, 0);
}
DEFUN ("re-search-backward", Fre_search_backward, Sre_search_backward, 1, 4,
@@ -2393,7 +2480,7 @@ since only regular expressions have distinguished subexpressions. */)
int some_nonuppercase_initial;
register int c, prevc;
int sub;
- int opoint, newpoint;
+ EMACS_INT opoint, newpoint;
CHECK_STRING (newtext);
@@ -2436,7 +2523,7 @@ since only regular expressions have distinguished subexpressions. */)
if (NILP (fixedcase))
{
/* Decide how to casify by examining the matched text. */
- int last;
+ EMACS_INT last;
pos = search_regs.start[sub];
last = search_regs.end[sub];
@@ -2523,8 +2610,8 @@ since only regular expressions have distinguished subexpressions. */)
if desired. */
if (NILP (literal))
{
- int lastpos = 0;
- int lastpos_byte = 0;
+ EMACS_INT lastpos = 0;
+ EMACS_INT lastpos_byte = 0;
/* We build up the substituted string in ACCUM. */
Lisp_Object accum;
Lisp_Object middle;
@@ -2664,7 +2751,7 @@ since only regular expressions have distinguished subexpressions. */)
/* Note that we don't have to increment POS. */
c = SREF (newtext, pos_byte++);
if (buf_multibyte)
- c = unibyte_char_to_multibyte (c);
+ MAKE_CHAR_MULTIBYTE (c);
}
/* Either set ADD_STUFF and ADD_LEN to the text to put in SUBSTED,
@@ -2686,7 +2773,7 @@ since only regular expressions have distinguished subexpressions. */)
{
c = SREF (newtext, pos_byte++);
if (buf_multibyte)
- c = unibyte_char_to_multibyte (c);
+ MAKE_CHAR_MULTIBYTE (c);
}
if (c == '&')
@@ -2714,7 +2801,7 @@ since only regular expressions have distinguished subexpressions. */)
set up ADD_STUFF and ADD_LEN to point to it. */
if (idx >= 0)
{
- int begbyte = CHAR_TO_BYTE (search_regs.start[idx]);
+ EMACS_INT begbyte = CHAR_TO_BYTE (search_regs.start[idx]);
add_len = CHAR_TO_BYTE (search_regs.end[idx]) - begbyte;
if (search_regs.start[idx] < GPT && GPT < search_regs.end[idx])
move_gap (search_regs.start[idx]);
@@ -2768,9 +2855,9 @@ since only regular expressions have distinguished subexpressions. */)
/* Adjust search data for this change. */
{
- int oldend = search_regs.end[sub];
- int oldstart = search_regs.start[sub];
- int change = newpoint - search_regs.end[sub];
+ EMACS_INT oldend = search_regs.end[sub];
+ EMACS_INT oldstart = search_regs.start[sub];
+ EMACS_INT change = newpoint - search_regs.end[sub];
int i;
for (i = 0; i < search_regs.num_regs; i++)
@@ -2849,7 +2936,7 @@ DEFUN ("match-data", Fmatch_data, Smatch_data, 0, 3, 0,
Element 2N is `(match-beginning N)'; element 2N + 1 is `(match-end N)'.
All the elements are markers or nil (nil if the Nth pair didn't match)
if the last match was on a buffer; integers or nil if a string was matched.
-Use `store-match-data' to reinstate the data in this list.
+Use `set-match-data' to reinstate the data in this list.
If INTEGERS (the optional first argument) is non-nil, always use
integers \(rather than markers) to represent buffer positions. In
@@ -3026,7 +3113,7 @@ If optional arg RESEAT is non-nil, make markers on LIST point nowhere. */)
}
else
{
- int from;
+ EMACS_INT from;
Lisp_Object m;
m = marker;
@@ -3194,20 +3281,20 @@ syms_of_search ()
}
searchbuf_head = &searchbufs[0];
- Qsearch_failed = intern ("search-failed");
+ Qsearch_failed = intern_c_string ("search-failed");
staticpro (&Qsearch_failed);
- Qinvalid_regexp = intern ("invalid-regexp");
+ Qinvalid_regexp = intern_c_string ("invalid-regexp");
staticpro (&Qinvalid_regexp);
Fput (Qsearch_failed, Qerror_conditions,
- Fcons (Qsearch_failed, Fcons (Qerror, Qnil)));
+ pure_cons (Qsearch_failed, pure_cons (Qerror, Qnil)));
Fput (Qsearch_failed, Qerror_message,
- build_string ("Search failed"));
+ make_pure_c_string ("Search failed"));
Fput (Qinvalid_regexp, Qerror_conditions,
- Fcons (Qinvalid_regexp, Fcons (Qerror, Qnil)));
+ pure_cons (Qinvalid_regexp, pure_cons (Qerror, Qnil)));
Fput (Qinvalid_regexp, Qerror_message,
- build_string ("Invalid regexp"));
+ make_pure_c_string ("Invalid regexp"));
last_thing_searched = Qnil;
staticpro (&last_thing_searched);
@@ -3239,6 +3326,8 @@ is to bind it with `let' around a small expression. */);
defsubr (&Ssearch_backward);
defsubr (&Sword_search_forward);
defsubr (&Sword_search_backward);
+ defsubr (&Sword_search_forward_lax);
+ defsubr (&Sword_search_backward_lax);
defsubr (&Sre_search_forward);
defsubr (&Sre_search_backward);
defsubr (&Sposix_search_forward);